1. Field of the Invention
[0001] The present invention relates to a detection method for detecting knocking in an
internal combustion engine and an ignition timing control method.
2. Description of Prior Art
[0002] As is well known, knocking is a phenomenon resulting from a self-ignition of unburned
gas remaining in the extremity region of the combustion chamber, and causing vibration
in the gas inside the combustion chamber, consequently propagating thus generated
vibration to the engine body itself.
[0003] Since this knocking gives rise to a loss of energy in the engine output, exerts mechanical
shock on the components of the engine as well as lowers the fuel consumption and the
like, it is desired to be avoided as much as possible. For this purpose, it is indispensable
for a precise detection of the occurrence of a knocking to be ensured to be made.
[0004] Reflecting such requirements, there has been proposed a prior art as set forth in
the Japanese Patent Laid-open No.58-45520/1983 wherein the occurrence of knocking
is detected by going through the steps of separating resonance frequency components
in the range of 5 to 20 kHz in the output signals from the vibration detection sensor
with band pass filters, and determining whether a value is greater than the background
level thereof.
[0005] However, such knocking detection method based on single resonance frequency components
involves such problems that the background level thereof becomes larger at a faster
engine speed and that the knocking resonance frequencies tend to change with varying
engine data.
DESCRIPTION OF THE INVENTION
[0006] Therefore, in order to solve the problems associated with the prior art, there has
been proposed a knock detection method as set forth in the Japanese Patent Laid-open
No.3-47449/1991 (EP-A2,3-392 804) wherein the detection of knocking is carried out
by sampling a plurality of resonance frequency components.
[0007] This knock detection method has been very innovative and successful in realizing
a high precision knock detection, and thus has been highly appraised and accepted
by the related industry.
[0008] Reflecting recent demands for further improvements in fuel consumption, exhaust emission
cleaning capabilities and the like, many attempts to increase the compression ratio
in the engine are under way. Increased compression ratios, however, in turn tend to
cause a knocking readily to occur, thereby demanding further efforts to improve the
knock detecting precision.
[0009] However, for an engine having a poor frequency characteristic, there exist such operational
conditions and frequency bands which cause the background level thereof to temporarily
or continuously change, for example, to decrease, depending on the performance of
the vibration sensors employed, or positions of attachment thereof and the like.
[0010] Therefore, when there occurs that, for example, a background level utilized as a
parameter for detecting the occurrence of knocking becomes abnormally small, there
may be formed a false signal indicating an occurrence of knocking in at least one
of the frequency bands of the plurality of resonance frequencies, notwithstanding
that actually no knocking is caused therein.
[0011] Thereby, when there is formed a false signal indicating an occurrence of knocking
at least in one of the frequency bands, nevertheless, in the plurality of resonance
frequencies to be sampled for a precise detection of knocking, this false signal adversely
affects the whole knocking determination performance over these resonance frequencies,
consequently impeding further to improve the knocking detection precision.
[0012] According to a first aspect of this invention there is provided a knocking detection
method as claimed in claim 1 herein.
[0013] According to a further aspect of this invention there is provided an ignition timing
control method as claimed in claim 19 herein.
[0014] Since when any comparison component, i.e., background level varies abnormally a limit
comparison component is adapted to be used instead of the abnormally varying comparison
component, information on knocking is ensured to be obtained from the relationship
between the limit comparison component and the characteristic component as well, thereby
eliminating an issuance of abnormal information on knocking, thus improving the detecting
precision of knocking.
BRIEF DESCRIPTION OF THE DRAWINGS
[0015] The invention will be described more particularly in the following with reference
to the accompanying drawings, in which;
Fig.1 is a flowchart illustrating a knocking detection method of one embodiment of
the invention;
Fig.2 is a vibration waveform diagram without any knock generation;
Fig.3 is a vibration waveform diagram with a knock generation;
Fig.4 illustrates spectrum intensity of a knocking disclosed in EP-A-0 392 804;
Fig.5 illustrates another spectrum intensity of a knocking disclosed in EP-A-0 392
804;
Fig.6 illustrates power spectra with and without any knock generation;
Fig.7 is a vibration waveform diagram;
Fig.8 is a background level diagram;
Fig.9 shows a knock determination diagram;
Fig.10 shows lower limiter maps;
Fig.11 is a flowchart for reading out a lower limiter;
Fig.12 illustrates a calculation method for calculating a lower limiter;
Fig.13 is a flowchart illustrative of arithmetic operation of a lower limiter;
Fig.14 shows a system configuration of the invention;
Fig.15 shows a schematic block diagram of a control unit of the invention;
Fig.16 is a flowchart illustrating ignition calculation according to the invention;
Fig.17 shows a lower limiter map;
Fig.18 is a flowchart for reading out a lower limiter;
Fig.19 shows a lower limiter map;
Fig.20 is a flowchart for reading out a lower limiter;
Fig.21 shows a lower limiter map;
Fig.22 is a flowchart for reading out a lower limiter;
Fig.23 is a flowchart illustrating another embodiment of the invention;
Fig.24 is a diagram illustrative of switching between lower limiters;
Fig.25 is a chart diagram explaining acceleration steps;
Fig.26 is a flowchart illustrating still another embodiment of the invention; and
Fig.27 is a flowchart illustrating further embodiment of the invention.
DESCRIPTION OF PREFERRED EMBODIMENTS
[0016] With reference to the accompanying drawings, one embodiment of the invention will
be set forth in detail in the following.
[0017] At first, the knocking determination principle of the invention for determining an
occurrence of a knocking will be explained. There are contained various vibration
components in the vibration of an engine. Such vibration components include, for example,
those caused by piston friction, crank shaft rotation, valve operation and the like.
In addition, these vibrational components tend to vary responsive to engine conditions.
[0018] When knocking occurs in the engine, a characteristic vibration specific to knocking
is generated. Thereby, determination of a knocking occurrence is realized by separating
the characteristic vibration inherent to knocking from the whole vibration components
of the engine detected by the vibration sensor.
[0019] Figure 2 shows a result of frequency analysis of an output frequency component from
a vibration sensor when no knocking is present. On the other hand, Figure 3 shows
a result of frequency analysis of an output frequency component from the vibration
sensor when knocking is present.
[0020] As is obviously understood from the comparison of Figures 2 and 3, in the case where
knocking is present, respective resonance frequency components tend to become larger
than in the case where no knocking is present.
[0021] Next, with reference to Figures 4 and 5, a knocking occurrence determination method
utilizing a knocking determination vector index will be described in the following.
The operational principle of the invention will be described by way of example of
resonance frequency components f
10(6.3 kHz) and f
01(13.0 kHz). They are, however, not limited thereto, and at least two of any resonance
frequency components may be utilized for determining a knocking occurrence.
[0022] The vibration sensor detects a vibration which has been synthesized by combining
a vibration component due to a knocking occurrence and one due to a background thereof.
Thereby, when there is no knocking a knocking determination index I becomes an index
I
b which is defined by the background vibration, while when there is knocking it becomes
an index I which is defined so as to include the background vibration I
b and a vibration component I
k due to the occurrence of knocking.
[0023] The hereabove knocking determination index I can be formularized into the following
equation including major resonance frequency components.

where, ω is a real value to be determined by an engine speed. Further, it may take
a binary value of 1 or 0. P represents power spectra of respective resonance frequency
components.
[0024] As shown in Fig.4, the knock determination vector index I
b which is expressed by the resonance frequency components of the background vibrations
and the index I
k which is expressed by the resonance frequency components of the vibration resulting
from the occurrence of a knocking have a different direction and magnitude from each
other. This is because that, as in the audible test in which the occurrence of a knocking
is identified by a crunching noise, there is involved a change in noise tone when
a knocking occurs.
[0025] When the vibration component due to the occurrence of a knocking is added to the
background vibration, the knock determination index I according to f
01, f
10 included in the original from the vibration sensor falls in a range below a knock
determination threshold value I
01 in Fig.4, and outside a dotted arch line of a threshold value I
02 in Fig.5, thus enabling determination of the occurrence of knocking.
[0026] Further, in the present invention, any set of the plurality of resonance frequency
components included in the output from the vibration sensor and utilized in combination,
not limiting to the five elements on the right-hand term of Equation (1), is prescribed
to be a knocking determination index.
[0027] When such a knocking determination index is utilized, the occurrence of any knocking
is ensured to be determined even if a background vibration increases significantly
since the composition of a characteristic frequency component indicative of a knocking
relative to its corresponding background vibration is taken into consideration.
[0028] As set forth previously, there are some operational conditions and frequency bands
where the background level decreases temporarily or continuously depending on the
quality of the vibration sensors employed or their attachment positions, consequently
causing a false signal indicative of the occurrence of a false knocking notwithstanding
actually no knocking being present.
[0029] When attempting to obtain energy of the knocking resonance frequencies on the basis,
for example, of a signal/noise (background) ratio, i.e., S/N ratio, an abnormal decrease
in the background level in one of the resonance frequencies will cause its S/N ratio
to become abnormally larger, which eventually affects the total determination processing
thereof on the basis of the plural resonance frequencies, in effect curbing the efforts
to improve the knocking detection precision.
[0030] In order to solve the hereinabove problems associated with the prior art, when a
particular background level corresponding to at least one of the resonance frequencies
decreases abnormally, the present invention is adapted not to utilize the abnormally
decreased background level but to utilize a predetermined limit value, for example,
background lower limiter.
[0031] In the following, a knocking detection method embodying the invention constituting
a key portion thereof will be described in detail.
[0032] Figure 6 shows power spectra relative to the frequencies, in which the occurrence
of a knocking is indicated by a curve represented by a solid line, while the non-occurrence
thereof is represented by broken lines. It is clearly shown that power spectra of
respective resonance frequency bands are increased due to the occurrence of knocking.
[0033] However, in case there exist characteristic frequencies inherent to a knocking having
smaller power spectra as shown by f
1 and f
9, or in case there exist greater noise components depending on the nature of the engine
as shown by A in Figure 7, there occurs such a phenomenon as shown by B in Figure
8 in which the background level is caused to drop abnormally.
[0034] Thereby, there has been a problem that an S/N ratio becomes abnormally large, eventually
causing a final knocking determination index to be deviated, thereby giving rise to
an erroneous judgment.
[0035] Thereby, it is arranged according to the invention to eliminate erroneous judgment
on knocking by setting a lower limiter as shown by broken lines with respect to the
background level in Figure 8.
[0036] First, with reference to Figure 1, the operation of a knocking occurrence determination
process in a CPU according to the invention will be described in the following.
[0037] The flowchart in Figure 1 is executed for every detonation per cycle, and is activated
by interrupting the CPU.
[0038] In step 101, an output signal from the vibration sensor is read after its analog
to digital conversion in an A/D converter.
[0039] In the next step 102, a frequency analysis of the analog-to-digital converted signal
from the vibration sensor is executed. This frequency analysis is carried out through
a transform method such as the fast Fourier transform or Walsh transform.
[0040] Then, in step 103, a plurality of frequency bands each containing a resonance frequency
are selected from the signals which have been subjected to the frequency analysis.
In this embodiment of the invention, a total of eight resonance frequency bands are
selected.
[0041] Upon selection of preferred frequencies at step 103, in the next step 104 an S/N
ratio indicative of a power spectrum is obtained for each selected frequency.
[0042] Namely, there are obtained a plurality of selected frequencies (f
1 ...... f
i), i.e., f
1 ...... f
8 in this embodiment, and a plurality of corresponding background levels (BGL
1 ...... BGL
i), i.e., BGL
1 ...... BGL
8 in this embodiment, then an S/N ratio SL
i=f
i/BGL
i is obtained for each frequency selected.
[0043] Therefore, in this embodiment of the invention, there are obtained respectively,

[0044] In the next step 105, a knock strength is obtained by sampling m numbers, in this
embodiment 5, from the selected frequencies which are ordered in the decreasing order
of S/N ratios. An equation to obtain this knock strength is expressed for example
as follows, whereby the S/N ratio is obtained through adding operation.

[0045] When the knock strength is obtained in step 105 it is compared with a predetermined
value obtained in step 106 for knock determination. When it is judged that the knock
strength obtained in step 105 is larger than the predetermined value, the occurrence
of a knocking is identified in step 107.
[0046] Then, a knock flag "1" indicative of the occurrence of a knocking is set in step
108. This knock flag is utilized in an ignition control task which is activated separately.
[0047] On the other hand, when the knock strength is judged to be smaller than the predetermined
value at step 106, assuming that no knocking is taking place, it is judged in step
109 whether each background level BGL
i is larger than a predetermined limit value, namely in this embodiment, lower limiter
BGLMT
i. Thereby, in this embodiment of the invention, BGL
1 ...... BGL
8 are compared with corresponding BGLMT
1 ...... BGLMT
8, respectively.
[0048] When a background level is judged in step 109 to be greater than the lower limiter
BGLMT
i, that is, to be a normal background level, the background level BGL
i is updated in step 110.
[0049] An update value for this background level BGL
i is obtained by filter processing of the power spectrum of a selected frequency. More
specifically, it is obtained for each one of the selected frequencies by calculating
the following equation,

[0050] In contrast, when the background level BGL
i is judged in step 109 to be smaller than the lower limiter BGLMT
i, that is, the background level BGL
i decreases abnormally, the lower limiter BGLMT
i is set instead thereof in step 111 to be utilized as a BGL
i in the next step 104 to follow.
[0051] Then, in step 112, the knock flag is set to "0".
[0052] Through the hereabove processing the knock detection routine is accomplished, then
the knock flag set in this routine is utilized in the ignition control task. In addition,
although the lower limiter herein is also utilized as a threshold value for background
level determination, they may be provided separately.
[0053] The lower limiter BGLMTi will be described in detail in the following.
[0054] Various modifications of the lower limiter BGLMT
i of the invention may be considered, but the most typical examples of which will be
such methods: one is by storing data; the other is through calculation thereof. They
will be described more specifically in the following.
[0055] With reference to Figure 10, a method for setting of a lower limiter BGLMT
i according to mapping is shown, where it is set for respective engine speeds such
as 1000 rpm, 2000 rpm, 3000 rpm in this embodiment, and corresponding to respective
cylinders (from 1 to 6 cylinders in this example) and respective resonance frequencies
(f
o ······ f
i).
[0056] By adopting such mapping according to the invention, it is possible to set a very
high-precision BGLMT
i.
[0057] Further, these three parameters of the engine speeds, cylinder numbers and frequencies
cited hereabove are not limited thereto, and it should be construed that it is also
possible to set an appropriate BGLMT
i with at least one of these parameters. A preferred result has been obtained, however,
when the BGLMT
i is set at least for respective frequencies. As a matter of course, it is possible
for any BGLMT
i stored in the map in Figure 10 to be retrieved through step 113 in Figure 11 by referring
to a corresponding engine speed, cylinder number, and frequency. The step 113 shown
in Figure 11 is inserted for execution thereof between steps 106 and 109 in Figure
1.
[0058] On the other hand, with reference to Figures 12 and 13, the method for obtaining
BGLMT
i by calculation will be described in the following.
[0059] Figure 12 is a schematic diagram to help understand an underlying concept thereof.
A maximum value and a minimum value are obtained for each frequency thereby to derive
a mean value BGL
mean. Then, this mean value BGL
mean is divided into by n (n=integer) so as to provide a lower limiter BGLMT
i.
[0060] How to obtain these mean value BGL
mean and lower limiter BGLMT
i are described more specifically in steps 114 and 115 in Figure 13, where in step
114 a maximum value of the background level BGL
up and a minimum value of BGL
BP are added and divided by 2 for each frequency, then summed up from 1 to m (for example,
16 times), then divided by m to obtain the BGL
mean, next in step 115 thusly obtained BGL
mean is divided by n thereby to obtain the lower limiter BGLMT
i. As a matter of course, it is possible to obtain respective BGLMT
i corresponding to respective cylinder numbers, engine driving conditions as well.
[0061] Steps 114 and 115 in Figure 13 are inserted for execution between step 106 and step
109 in Figure 1.
[0062] By way of example, regarding a constant 1/n used in obtaining the lower limiter BGLMT
i from the mean value BGL
mean, a 1/3 is found to be able to provide the best BGLMT
i.
[0063] The knock signal indicative of the knock occurrence thus obtained is utilized in
the ignition task which will be described in detail in the following.
[0064] Figure 14 shows a system configuration diagram of the ignition apparatus. Air is
introduced through an intake port of an air cleaner 1, then passes through an air
duct 3, a throttle body 5 having a throttle valve, and an air supply tube 6 to be
admitted into the cylinders of an engine 7. An inlet air quantity is detected by a
hot wire air flow sensor 2 mounted on the air duct 3, and a detected signal is input
into a control unit 9.
[0065] On the other hand, fuel supplied from a fuel tank (not shown) is injected through
an injector 16, mixed with intake air in the intake passage, then supplied into the
cylinders of the engine 7. Air fuel mixture is compressed by the engine 7, ignited
by an ignition plug 15, then after explosion the exhaust gas is exhausted from an
exhaust pipe 8. An exhaust sensor 11 is mounted on the exhaust pipe 8, and a detected
signal detected thereby is input into the control unit 9.
[0066] A high voltage generated by an ignition coil 13 is distributed to each cylinder by
a distributor 14 to be supplied to each ignition plug 15. An engine speed is detected
by a crank angle sensor 12. The crank angle sensor 12 outputs a reference signal Ref
indicating an absolute position for every rotation thereof and a P
os signal indicating a position thereof which is shifted by a predetermined angle from
said absolute position. The Ref signal and the P
os signal are input into the control unit 9 as well. A vibration sensor 151 is mounted
on the engine 7 for detection of vibration thereof, and a detection signal therefrom
is input into the control unit 9.
[0067] The control unit 9 in response to signals from respective sensors calculates an optimum
fuel supply quantity, ignition timing and the like, then outputs a control signal
thus calculated to the injector 16 and the ignition coil 13 respectively.
[0068] Figure 15 shows in detail a configuration diagram of the control unit 9 of the invention.
The control unit 9 is divided into two blocks of a control block 34 and a knocking
detecting block 35: the control block 34 comprising a CPU 20, an A/D converter 21,
a ROM 22, an input I/O 23, a RAM 24, a DPRAM 25, an output I/O 26 and a bus 37; and
the knocking detecting block 35 comprising a CPU 29, a port 27, a timing circuit 28,
an A/D converter 30, a ROM 31, a RAM 32, a clock 33, an operational circuit 38 and
a bus 36. By way of example, exchange of data between the CPU 20 and CPU 29 is executed
through the DPRAM 25 which is a dual port RAM.
[0069] An intake air flow Q
a detected by the hot wire air flow meter 2 is converted into a digital value through
the A/D converter 21, then is input into the CPU 20. Further, a Ref signal and a P
os signal detected by the crank angle sensor 12 are input into the CPU 20 through the
input I/O 23. The CPU 20 executes arithmetic operation of thus entered signals according
to a program stored in the ROM 22, then the result of the arithmetic operation is
output through the output I/O 26 as a fuel injection time interval signal T
i which designates a fuel injection amount, and an ignition timing signal θ
ign to respective associated actuators. Data storage for storing data required in the
arithmetic operation is implemented by the RAM 24.
[0070] On the other hand, when the operational circuit 35 generates a top dead center (TDC)
signal, the timing circuit 28 in response to the contents of data input by the CPU
20 through the port 27 divides a frequency signal generated and supplied from the
clock 33, then generates on the basis of thus divided frequency signal a sampling
signal. Upon occurrence of the sampling signal, the A/D converter 30 converts an output
signal from the vibration sensor 15 into a digital value.
[0071] Prior art vibration sensors for use in detecting the knocking generally initiate
resonance in the vicinity of 13 KHz, however, in this embodiment of the invention,
in order to obtain resonance frequency components up to 18 to 20 KHz, such vibration
sensors that resonate above 18 KHz are utilized.
[0072] The CPU 29 stores the sampled digital value in the RAM 32 according to the program
stored in the ROM 31 as well as executes a frequency analysis thereof on the basis
of the data stored according to the flowchart shown in Fig.1, thereby enabling detection
of the occurrence of knocking taking place therein. The result of knocking occurrence
detection is transmitted to the CPU 20 through the DPRAM 25.
[0073] Next, with reference to a flowchart in Figure 16, an arithmetic operation for obtaining
an ignition timing in the CPU 20 will be described in the following. This flowchart
operation is activated at a predetermined time cycle, for example, at every 10 msec.
In step 201, an engine speed N and an intake air quantity Q are read out from predetermined
registers set in the RAM 24. In the next step 202, an intake air quantity per unit
engine speed Q/N is calculated, then from this Q/N is obtained a fuel injection time
width T
i, then a basic ignition timing θ
base is obtained from a set of basic ignition timing map stored in the ROM 22 to supply
fuel in accordance therewith. In step 203, it is judged whether a knocking has occurred
or not according to the contents of the knock flag set in the flowchart of Figure
1. If a knocking has occurred, a predetermined retard angle quantity Δθ
ret is subtracted from an ignition timing θ
adv in step 213. By this subtraction, the ignition timing is retarded. In step 214, the
retarded ignition timing retarded as the result of the knocking occurrence is compared
with a predetermined speed, for example, 50 (as indicated in step 205) to determine
the base for a value to be recovered. Count data A is initialized, then the step advances
to 208.
[0074] On the other hand, when there occurred no knocking in step 203, the count data A
is incremented by 1. The count data A is used to determine whether it is a time for
the retarded ignition timing θ
adv retarded due to the occurrence of knocking to be recovered by a leading angle quantity
Δθ
adv. In step 205, it is judged whether the count data A equals a predetermined value
50. Since the flowchart of Fig.16 is activated every 10 msec, it is 0.5 sec after
initialization of the counter data A when the count data A becomes equal to 50, thus
it is recovered every 0.5 sec. When the count data A is not equal to 50 in step 205,
the step advances to 206. In step 206, a predetermined leading angle quantity Δθ
adv is added to the retard angle value θ
adv. By this addition, the ignition timing is effectually recovered.
[0075] Next, in step 208, an ignition timing θ
ign is calculated by adding the ignition timing θ
adv hereinabove obtained to the basic ignition timing θ
base. In step 209, a maximum lead angle value θ
res is obtained corresponding to the engine speed N and the unit value Q/N of the intake
air quantity to engine speed. Obtaining of the maximum lead angle value θ
res is executed by reading from the maximum lead angle value map stored in the ROM 31.
In step 210 it is judged whether the ignition timing θ
ign has exceeded the maximum lead angle value θ
res or not. If not, the step advances to 211. If exceeded the maximum lead angle value
θ
res, the lead angle being too excessive, the maximum lead angle value θ
res is set as an ignition timing θ
ign in step 211.
[0076] Finally, after setting of the ignition timing θ
ign, a delay time t
d, a sampling number n
s and a frequency division ratio t
s corresponding to an engine state are output to the port 27 in step 212.
[0077] By way of example, according to the frequency division ratio t
s, a sampling cycle for sampling output digital values from the vibration sensor is
determined, and according to the sampling number n
s the number of sampling is determined.
[0078] As hereinabove set forth, through detection of knocking from the plurality of resonance
frequency components and by controlling the ignition timing appropriately responsive
thereto, the occurrence of knocking in the engine can be eliminated.
[0079] For the hereinabove knock occurrence signals to be utilized in the ignition control
task there is set basically for each frequency a lower limiter BGLMT
i as shown in Figs. 1, 10, 12. However, if any particular resonance frequency band
is known in advance in which the background level is anticipated to fall abnormally,
it is possible to set a lower limiter corresponding to such frequency band alone.
For example, with reference to Figure 6, if it is known in advance that background
levels corresponding to resonance frequencies f
1 and f
2 will fall, only the lower limiters BGLMT
1 and BGLMT
9 alone may be set.
[0080] Further, in case a full range of lower limiters BGLMT
i as in Figure 10 cannot be prepared because of the limits in terms of amounts of calculation
involved and map capacity required, the following alternate method may be adopted.
[0081] That is, a limited number of representative lower limiters BGLMTs corresponding to
typical resonance frequencies, representative cylinders or some typical engine driving
conditions may be selected to be set for use in common.
[0082] First, Figure 17 illustrates setting of lower limiters BGLMTs with respect to a particular
typical resonance frequency.
[0083] A map in Figure 17 is prepared corresponding both to respective cylinders and respective
engine speeds with respect to a particular resonance frequency f
i=6.3 KHz.
[0084] Thereby, a lower limiter BGLMT determined corresponding to a particular cylinder
number and a particular engine speed is used in common for comparison with each background
level BGL
i corresponding to the selected resonance frequency.
[0085] That is, in step 116 shown in Figure 18, a corresponding lower limiter BGLMT is retrieved
corresponding to a specific engine speed and a specific cylinder number from the map
of Figure 17. Then, in step 117, a comparison is carried out for each resonance frequency
BGL
i with the common lower limiter BGLMT. These two steps are inserted between steps 106
and 109 in Figure 1. As a matter of course, step 117 is intended to replace step 109.
Further, it should be construed that BGLMT
i in step 111 is a common lower limiter BGLMT.
[0086] Figure 19 shows an example of a map prepared corresponding to respective frequencies
and respective engine speeds with respect to a particular number of cylinders 3.
[0087] Thereby, a lower limiter BGLMT to be determined at a cross point corresponding to
a respective frequency and a respective engine speed is used for comparison with each
background level BGL
i of the selected resonance frequency.
[0088] Namely, as shown in Figure 20, a corresponding lower limiter BGLMT corresponding
to a specific engine speed and a specific frequency is retrieved from the map of Figure
19, then in step 119 a comparison of the corresponding lower limiter BGLMT with a
background level is carried out for respective resonance frequency. Thereby, since
it is used in common as the lower limiter for each cylinder, these steps are inserted
between steps 106 and 109 in Figure 1. Step 119 is intended to replace step 109.
[0089] Figure 21 shows an example of a map formulated corresponding to respective resonance
frequencies and respective cylinders, intended for use in a specific driving condition
of the engine. That is, the specific engine speed selected herewith is used as a parameter
in common.
[0090] Thereby, a lower limiter BGLMT to be determined at a position in coincidence corresponding
to a particular frequency and a particular cylinder number is used in common for comparison
with a respective background level BGL
i of the selected resonance frequency.
[0091] Namely, in step 120 in Figure 22, a corresponding lower limiter BGLMT corresponding
to a specific frequency and a specific cylinder number is retrieved from the map of
Figure 21, then in step 121 this corresponding lower limiter BGLMT is used for comparison
with a background level for respective resonance frequencies. Thereby, since it is
used in common as the lower limiter for the selected engine speed, these steps are
inserted between steps 106 and 109 in Figure 1. It should be construed, however, that
the step 121 replaces the step 109.
[0092] When it is required further to reduce the calculation loads and mapping memory capacities,
it is also possible to set up the best mode of a lower limiter BGLMT obtainable in
the most desirable conditions with respect to the resonance frequencies, cylinder
numbers and engine speeds, respectively.
[0093] That is, only one lower limiter BGLMT is effected to be set irrespective of respective
resonance frequencies, cylinder numbers and engine speeds. Thereby, this only one
lower limiter is stored alone in a ROM area.
[0094] With reference to Figure 23, another embodiment of the invention will be set out
in the following. Since the basic configuration thereof is the same as in Figure 1,
only the gist thereof will be described.
[0095] In Figure 23, step 122 is executed when it is judged in step 106 that no knocking
has occurred, and in step 122 the singular lower limiter BGLMT thus set hereinabove
irrespective of respective resonance frequencies, cylinder numbers, engine speeds
is effected to be compared with a background level BGL
i corresponding to each resonance frequency. When it is judged as a result of the comparison
that any background level BGL
i is smaller than the lower limiter BGLMT, the lower limiter BGLMT is set up in step
123 instead thereof so as to ensure that in the next recurring step 104 the BGL
i is replaced by the BGLMT and each S/N ratio is obtained thereby for each frequency.
[0096] Further, the concepts and methods of the invention described in Figures 17 through
23 may be implemented by calculations as well, according to the calculational method
as specified in Figure 13.
[0097] Still further, a method for obtaining the background level BGL
i in the various embodiments of the invention described heretofore is based on the
following equation for applying a filter processing to a signal intensity f
i,

However, it may be obtained by the usual integral method of integrating background
level measurement as well.
[0098] Further, although the power spectrum SL
i is obtained from the S/N ratio, it may be obtained as well from a difference from
the background level BGL
i, i.e., SL
i=f
i-BGL
i. Thereby, a knock strength S may be obtained by adding up this difference.
[0099] Still further, although it is judged that there occurred knocking when an added value
of the power spectra of the plurality of frequencies exceeds a preset value, there
may be provided a logic which prescribes the occurrence of a knocking if any one of
the power spectra of the plurality of frequencies exceeds a preset value, without
going through addition of respective power spectra.
[0100] In the next, there will be described a compensation method for compensating for a
side effect occurring when the background level BGL
i is obtained through the filter processing thereof.
[0101] Since the background level BGL
i in the one embodiment of the invention shown in Figure 1 is obtained through the
delay filter processing, there arises a problem in a transient state of the engine
such as acceleration or the like that the background level BGL
i cannot follow up the transient state and thereby that an apparent S/N increases.
[0102] The lower limiter BGLMT
i according to the present invention is effective also to deal with such follow-up
delays of BGL
i as well, but it is further advantageous to change the lower limiter BGLMT
i according to acceleration and steady state operation.
[0103] Figure 24 shows transient-state and steady-state lower limiters BGLMT
i, where the transient-state limiter is set at a larger value.
[0104] Figure 25 shows changes, in a transient state, of the throttle valve opening, engine
speeds, background level BGL
i and lower limiter BGLMT
i, wherein it is judged according to a degree of the throttle valve opening per unit
period of time whether is a rapid acceleration or a slow acceleration, and depending
on either it is a rapid acceleration or a slow acceleration a magnitude of the lower
limiter BGLMT
i is adjusted. This adjustment is effected either by multiplying the lower limiter
BGLMT
i for use in the steady-state operation by a constant or adding thereof.
[0105] The foregoing flowchart will be described with reference to Figure 26, where in step
124 a degree of throttle opening is determined in what level it is; if it is a rapid
acceleration, intermediate acceleration, slow acceleration or any other driving condition.
Then, when it is judged to be a rapid acceleration in step 125, the steady-state lower
limiter BGLMT
i is multiplied by a coefficient 1.5, when it is judged to be an intermediate acceleration
in step 126 the steady-state lower limiter BGLMT
i is multiplied by a coefficient 1.3, when it is judged to be a slow acceleration in
step 127 the steady-state lower limiter BGLMT
i is multiplied by a coefficient 1.1, and when it is judged to be a steady coursing,
the lower limiter BGLMT
i is set as it is in step 128. These steps are inserted for execution between steps
106 and 109 in Fig.1.
[0106] Through such arrangement, an appropriate lower limiter BGLMT
i can be obtained for various driving conditions of the engine.
[0107] In the embodiments of the invention described heretofore, attention has been directed
to such cases where the background level abnormally drops. However, there may also
take place such a phenomenon where the background level abnormally rises.
[0108] In such a case as above, an effective countermeasure can be provided by setting up
an upper limiter instead of the lower limiter likewise in step 109 in the foregoing
embodiment of the invention, .
[0109] In this case also, it should be construed that the upper limiter can take a variety
of modifications of value likewise the case of the lower limiter.
[0110] Further, it will be more advantageous to provide for a countermeasure against such
inconveniences anticipated to take place as will be described in the following.
[0111] Namely, gain of the background level indicating an amplification degree thereof is
switched, for example, to 1/2 at a predetermined timing by means of hardware, and
in line with this switching the power spectrum of each frequency is switched to 1/2
by means of software.
[0112] However, when the power spectrum of each frequency happens to drop abnormally, although
the background gain is ensured to be switched by the hardware, there occurs sometimes
a problem that the power spectrum of each frequency fails in switching due to an error
in quantization, consequently increasing the S/N ratio abnormally, thereby causing
an erroneous knocking detection.
[0113] In order to solve such problems, the following steps are added between steps 103
and 105 in Figure 27 according to the invention.
[0114] First, in step 129 following step 103 it is judged for each frequency f
i whether its strength is larger than/ equal to a preset lower limiter fLMT
i, or not. If it is judged here to be larger/equal thereto, the step advances to 104
where to execute an ordinal S/N ratio calculation.
[0115] On the other hand, in step 129 following step 103 if it is judged for each frequency
f
i that its strength is smaller than a preset lower limiter fLMT
i, the step advances to 130 where an S/N ratio of the frequency spectrum f
i to the background level is set at "1". That is, there is executed a processing thereby
for assuming that no knocking is present with respect to the corresponding frequency.
[0116] Next in step 131, the corresponding frequency spectrum is replaced by the predetermined
lower limiter fLMT
i, then the step advances to 105 where a knock strength calculation is executed.
[0117] As hereinabove stated, the problem associated with the prior art that S/N ratio increases
abnormally due to the quantization error taking place when the intensity of each frequency
drops abnormally can be solved by setting a lower limiter instead of that erroneous
frequency intensity.
[0118] Also in this preferred embodiment of the invention, it should be construed that the
value of a lower limiter corresponding to each frequency may take a variety of modified
value likewise the foregoing embodiment of the background lower limiter.
[0119] The merits and advantageous effects according to the present invention will be summarized
as follows. Even if the background level is varied abnormally, because of the arrangement
of the invention, the abnormally changing background level is altered to a predetermined
limit value, thereby ensuring effective information indicative of a knocking occurrence
to be obtained.
[0120] Thereby, such problems can be solved that take place when notwithstanding that the
plurality of resonance frequencies are sampled for detection of the occurrence of
any knocking, if there happens an error signal indicating the occurrence of a false
knocking in at least one of the frequency bands, the error signal adversely affects
the general judgement on the knocking based on the resonance frequencies in general,
thereby in consequence remarkably improving reliability of knock detection apparatus.
1. Klopferkennungsverfahren zur Erkennung des Klopfzustandes in einer Brennkraftmaschine,
wobei das Verfahren die folgenden Schritte umfaßt:
Extrahieren einer Mehrzahl charakteristischer Komponenten (fi), die einen Klopfzustand
anzeigen, aus physikalischen Größen (BGLi), die sich auf einen Klopfzustand in der
Brennkraftmaschine beziehen, sowie einer Vergleichskomponente, die gemäß der genannten
physikalischen Größen definiert wird;
Extrahieren von Daten über einen Klopfzustand (107, 108, 112) aus einem Verhältnis
(106) zwischen der genannten Mehrzahl charakteristischer Komponenten und der genannten
Vergleichskomponente; und
Festlegen eines Grenzwertes (fLMTi, BGLMTi) bezüglich der genannten Mehrzahl von charakteristischen
Komponenten und der genannten Vergleichskomponente.
2. Verfahren nach Anspruch 1, wobei eine Mehrzahl von Vergleichskomponenten vorgesehen
ist, die gemäß der genannten Mehrzahl charakteristischer Komponenten auf der Basis
der genannten physikalischen Größen definiert werden;
wobei mindestens eine Vergleichskomponente aus der Mehrzahl von Vergleichskomponenten
durch eine vorbestimmte Grenzwert-Vergleichskomponente ersetzt wird, wenn der Wert
der entsprechenden Komponente außerhalb eines zulässigen Bereichs liegt.
3. Verfahren nach Anspruch 2, wobei festgestellt wird, ob alle Komponenten der genannten
Mehrzahl von Vergleichskomponenten innerhalb eines Bereichs zulässiger Werte liegen
oder nicht; und wobei
eine Grenzwert-Vergleichskomponente gemeinsam für alle Komponenten der Mehrzahl
von Vergleichskomponenten eingesetzt wird, die sich außerhalb des Bereichs zulässiger
Werte befinden.
4. Verfahren nach Anspruch 3, wobei die genannte Substitution durchgeführt wird, wenn
festgestellt wird, daß die genannte Vergleichskomponente kleiner ist als die genannte
Grenzwert-Vergleichskomponente.
5. Verfahren nach Anspruch 1, wobei der Schritt des Extrahierens einer Mehrzahl charakteristischer
Komponenten folgendes umfaßt:
Ausführen einer Frequenzanalyse einer Schwingung, die einen Klopfzustand einer Brennkraftmaschine
darstellt und den Schritt des
Extrahierens von Daten über den Klopfzustand aus dem Verhältnis zwischen einer Mehrzahl
charakteristischer Frequenzkomponenten, die einen Klopfzustand anzeigen, die aus den
Frequenzen der Frequenzanalyse ausgewählt werden, und einer Vergleichskomponente,
die durch die Ausführung eines vorbestimmten Filterverfahrens an der charakteristischen
Frequenzkomponente gewonnen wird, und wobei der genannte Begrenzungs- bzw. Einschränkungsschritt
folgendes umfaßt:
Ersetzen der genannten Vergleichskomponente durch eine Grenzwert-Vergleichskomponente,
wenn die genannte Vergleichskomponente kleiner ist als die genannte Grenzwert-Vergleichskomponente.
6. Verfahren nach Anspruch 5, wobei
die genannte Vergleichskomponente entsprechend im Verhältnis zu der genannten Mehrzahl
charakteristischer Frequenzkomponenten festgelegt wird; und wobei
die genannte Grenzwert-Vergleichskomponente ebenfalls entsprechend im Verhältnis zu
der genannten Mehrzahl charakteristischer Frequenzkomponenten festgelegt wird.
7. Verfahren nach Anspruch 5, wobei die genannte Vergleichskomponente und die genannte
Grenzwert-Vergleichskomponente für die entsprechenden Zylinder des Motors getrennt
festgelegt werden können.
8. Verfahren nach Anspruch 5, wobei die genannte Vergleichskomponente und die genannte
Grenzwert-Vergleichskomponente für entsprechende bestimmte Motorantriebszustände getrennt
festgelegt werden können.
9. Verfahren nach Anspruch 5, wobei die genannte Grenzwert-Vergleichskomponente gemäß
den Motorantriebszuständen verändert wird.
10. Verfahren nach Anspruch 9, wobei die genannte Grenzwert-Vergleichskomponente in einem
Beschleunigungszustand des Motors auf einen größeren Wert eingestellt wird als sonst.
11. Verfahren nach einem der Ansprüche 1 bis 5 und Anspruch 9, wobei die genannte Grenzwert-Vergleichskomponente
einen vorbestimmten und in einer Speichereinrichtung gespeicherten Wert verwendet.
12. Verfahren nach einem der Ansprüche 1 bis 5 und Anspruch 9, wobei die genannte Grenzwert-Vergleichskomponente
durch arithmetische Operation der genannten Vergleichskomponenten gewonnen wird.
13. Verfahren nach Anspruch 1, wobei eine Mehrzahl von Vergleichskomponenten vorgesehen
wird, die gemäß der genannten Mehrzahl physikalischer Größen definiert werden, die
der genannten Mehrzahl charakteristischer Komponenten entsprechen;
wobei festgestellt wird, ob sich eine der Komponenten der genannten Mehrzahl charakteristischer
Komponenten außerhalb des genannten zulässigen Wertebereichs der Komponenten befindet
oder nicht; und wobei
eine vorbestimmte charakteristische Grenzwert-Komponente eine Komponente aus der Mehrzahl
charakteristischer Komponenten ersetzt, wenn in bezug auf diese Komponente festgestellt
wird, daß sie sich außerhalb des zulässigen Wertebereichs für die Komponenten befindet.
14. Verfahren nach Anspruch 1, wobei eine Mehrzahl von Vergleichskomponenten vorgesehen
wird, die auf der Basis der genannten Mehrzahl physikalischer Größen definiert werden,
die der genannten Mehrzahl charakteristischer Komponenten entsprechen;
wobei festgestellt wird, ob sich alle Komponenten der genannten Mehrzahl charakteristischer
Komponenten innerhalb des genannten zulässigen Wertebereichs der Komponenten befinden
oder nicht; und wobei
eine vorbestimmte charakteristische Grenzwert-Komponente eine Komponente aus der Mehrzahl
charakteristischer Komponenten ersetzt, die sich außerhalb des zulässigen Wertebereichs
für die Komponenten befindet.
15. Verfahren nach Anspruch 1, wobei eine Mehrzahl von Vergleichskomponenten vorgesehen
wird, die auf der Basis der genannten Mehrzahl physikalischer Größen definiert werden,
die der genannten Mehrzahl charakteristischer Komponenten entsprechen;
wobei festgestellt wird, ob sich alle Komponenten der genannten Mehrzahl charakteristischer
Komponenten innerhalb des genannten zulässigen Grenzwertes befinden oder nicht; und
wobei
eine vorbestimmte charakteristische Grenzwert-Komponente eine Komponente aus der Mehrzahl
charakteristischer Komponenten ersetzt, die sich außerhalb des zulässigen Grenzwertes
befindet.
16. Verfahren nach Anspruch 1, wobei der genannte Schritt der Festlegung des Grenzwertes
den Ersatz der genannten Mehrzahl charakteristischer Komponenten durch eine charakteristische
Grenzwert-Komponente umfaßt, wenn die genannte charakteristische Komponente kleiner
ist als die charakteristische Grenzwert-Komponente.
17. Verfahren nach Anspruch 1, wobei der genannte Schritt der Festlegung des Grenzwertes
den Ersatz einer charakteristischen Grenzwert-Komponente für jede charakteristische
Komponente der Mehrzahl charakteristischer Komponenten umfaßt, wenn eine der Komponenten
der Mehrzahl charakteristischer Komponenten kleiner ist als die genannte charakteristische
Grenzwert-Komponente.
18. Verfahren nach Anspruch 1, wobei festgestellt wird, ob alle Komponenten der genannten
Mehrzahl charakteristischer Komponenten kleiner ist als die entsprechende charakteristische
Grenzwert-Komponente oder nicht; und wobei
eine entsprechende charakteristische Grenzwert-Komponente eine charakteristische
Komponente ersetzt, für die festgestellt worden ist, daß sie kleiner ist als die genannte
entsprechende charakteristische Grenzwert-Komponente.
19. Verfahren zur Steuerung der Zündzeitpunktverstellung, wobei das Verfahren die folgenden
Schritte umfaßt:
Extrahieren einer Mehrzahl charakteristischer Komponenten (fi), die einen Klopfzustand
anzeigen, aus physikalischen Größen (BGLi), die sich auf einen Klopfzustand in der
Brennkraftmaschine beziehen, sowie einer Mehrzahl von Vergleichskomponenten, die gemäß
der genannten physikalischen Größen definiert werden;
Extrahieren von Daten über das Auftreten eines Klopfzustandes (107, 108, 112) aus
einem Verhältnis (106) zwischen der genannten Mehrzahl charakteristischer Komponenten
und der genannten Mehrzahl von Vergleichskomponenten;
Einstellen der Zündzeitpunktverstellung, so daß ein Klopfen bei dessen Auftreten unterdrückt
wird;
Aktualisieren einer der genannten Vergleichskomponenten wenn kein Klopfen vorliegt
durch einen neuen Wert, der sich innerhalb eines Wertebereichs befindet, der für die
genannte Vergleichskomponente zulässig ist; und
Aktualisieren einer der genannten Vergleichskomponenten mit einer vorbestimmten Grenzwert-Vergleichskomponente,
wenn sich die genannte Vergleichskomponente außerhalb des zulässigen Wertebereichs
für die Komponenten befindet.
20. Steuerung der Zündzeitpunktverstellung nach Anspruch 19, wobei die Schritte des Einstellens
und Aktualisierens die folgenden Schritte umfassen:
Feststellen des Auftretens eines Klopfzustandes, wenn eine Summe von mindestens zwei
der genannten Komponenten größer ist als ein vorbestimmter Wert;
Einstellen der Zündzeitpunktverstellung, so daß ein Klopfen bei dessen Auftreten unterdrückt
wird, und zwar gemäß der Feststellung des obengenannten Auftretens des Klopfens;
Bestimmen, nach der Feststellung, daß kein Klopfen existiert, wenn die Summe von mindestens
zwei Komponentenverhältnissen kleiner ist als der genannte vorbestimmte Wert, ob die
genannten entsprechenden Vergleichskomponenten größer sind als entsprechende Grenzwert-Vergleichskomponenten,
oder ansonsten als eine bestimmte ausgewählte Grenzwert-Vergleichskomponente;
Aktualisieren der entsprechenden Vergleichskomponenten mit entsprechenden neuen Werten,
wenn diese größer sind als die Grenzwert-Vergleichskomponente; und
Aktualisieren der genannten entsprechenden Vergleichskomponenten mit entsprechenden
Grenzwert-Vergleichskomponenten, die diesen entsprechen, oder mit der genannten einen
Grenzwert-Vergleichskomponente, wenn die entsprechenden Vergleichskomponenten kleiner
sind als die Grenzwert-Vergleichskomponente(n).
1. Procédé de détection de cognement dans un moteur à combustion interne, comprenant
les étapes consistant à :
extraire une pluralité de composantes caractéristiques (fi) indicatives d'un cognement
à partir de quantités physiques (BGLi) associées à un état de cognement dans le moteur
à combustion interne, et une composante de comparaison devant être définie en fonction
desdites quantités physiques ;
extraire une information relative au cognement (107, 108, 112) à partir d'une relation
(106) entre ladite pluralité de composantes caractéristiques et ladite composante
de comparaison ; et
régler une valeur limite (fLMTi, BGLMTi) sur l'une de ladite pluralité de composantes
caractéristiques et de ladite composante de comparaison.
2. Procédé selon la revendication 1, dans lequel il est prévu une pluralité de composantes
de comparaison devant être définies comme correspondant respectivement à ladite pluralité
de composantes caractéristiques sur la base desdites quantités physiques,
une composante de comparaison limite prédéterminée est substituée à au moins l'une
de la pluralité de composantes de comparaison lorsque la valeur de cette composante
est à l'extérieur d'une gamme admissible.
3. Procédé selon la revendication 2, dans lequel une évaluation est faite pour savoir
si l'ensemble de ladite pluralité de composantes de comparaison se situe ou non en-deçà
d'une limite d'une valeur admissible ; et
une composante de comparaison limite prédéterminée est substituée pour être utilisée
en commun pour n'importe quelle composante de comparaison qui est évaluée comme étant
en dehors de la limite de la valeur admissible.
4. Procédé selon la revendication 3, dans lequel ladite substitution est exécutée lorsque
ladite composante de comparaison est évaluée comme étant inférieure à ladite composante
de comparaison limite.
5. Procédé selon la revendication 1, dans lequel l'étape d'extraction d'une pluralité
de composantes caractéristiques comprend l'étape consistant à
exécuter une analyse de fréquences d'une vibration représentant un état de cognement
d'un moteur à combustion interne, et l'étape consistant à
extraire une information sur le cognement, qui est dérivée d'une relation entre une
pluralité de composantes de fréquences caractéristiques indicatives du cognement,
qui sont sélectionnées à partir des fréquences soumises à ladite analyse de fréquences,
et une composante de comparaison, qui est obtenue par application d'un procédé de
filtrage prédéterminé à ladite composante de fréquence caractéristique, et ladite
étape de limitation consiste à
substituer une composante de comparaison limite à ladite composante de comparaison
lorsque ladite composante de comparaison devient inférieure à ladite composante de
comparaison limite.
6. Procédé selon la revendication 5, dans lequel
ladite composante de comparaison est réglée de manière à correspondre respectivement
à ladite pluralité de composantes de fréquences caractéristiques ; et
ladite composante de comparaison limite est également réglée de manière à correspondre
respectivement à ladite pluralité de composantes de fréquences caractéristiques.
7. Procédé selon la revendication 5, dans lequel ladite composante de comparaison et
ladite composante de comparaison limite sont adaptées pour être réglées séparément
pour des cylindres respectifs du moteur.
8. Procédé selon la revendication 5, dans lequel ladite composante de comparaison et
ladite composante de comparaison limite sont adaptées pour être réglées séparément
pour des conditions respectives particulières de fonctionnement du moteur.
9. Procédé selon la revendication 5, dans lequel ladite composante de comparaison limite
est modifiée en fonction des conditions respectives de fonctionnement du moteur.
10. Procédé selon la revendication 9, dans lequel ladite composante de comparaison limite
est réglée, dans une condition d'accélération du moteur, à une valeur supérieure à
une valeur présente par ailleurs.
11. Procédé selon l'une quelconque des revendications 1 à 5 inclusivement et 9, dans lequel
ladite composante de comparaison limite utilise une valeur prédéterminée et est mémorisée
dans des moyens de mémoire.
12. Procédé selon l'une quelconque des revendications 1 à 5 inclusivement et 9, dans lequel
ladite composante de comparaison limite est obtenue au moyen d'une opération arithmétique
sur lesdites composantes de comparaison.
13. Procédé selon la revendication 1, dans lequel il est prévu une pluralité de composantes
de comparaison devant être définies conformément auxdites quantités physiques correspondant
à ladite pluralité respective de composantes caractéristiques ;
une évaluation est faite pour savoir si l'une quelconque de ladite pluralité de composantes
caractéristiques se situe ou non à l'extérieur d'une gamme admissible de valeurs ;
et
une composante caractéristique limite prédéterminée est substituée à ladite composante
parmi la pluralité de composantes caractéristiques lorsque cette composante est évaluée
comme étant située à l'extérieur de ladite gamme admissible de valeurs de cette caractéristique.
14. Procédé selon la revendication 1, dans lequel il existe une pluralité de composantes
de comparaison devant être définies sur la base desdites quantités physiques correspondant
respectivement à ladite pluralité de composantes caractéristiques ;
une évaluation est faite pour savoir si l'ensemble de ladite pluralité de composantes
caractéristiques se situe ou non dans une gamme admissible de la valeur correspondant
à ces composantes ; et
une composante caractéristique limite prédéterminée est substituée à l'une quelconque
des composantes caractéristiques qui est située à l'intérieur de la gamme admissible
de valeurs.
15. Procédé selon la revendication 1, dans lequel il est prévu une pluralité de composantes
de comparaison devant être définies sur la base desdites quantités physiques correspondant
respectivement à ladite pluralité de composantes caractéristiques ;
une évaluation est faite pour déterminer si l'ensemble de ladite pluralité de composantes
caractéristiques est inférieure ou non à une limite admissible de valeur ; et
une composante caractéristique limite prédéterminée est substituée à l'une quelconque
des composantes caractéristiques qui se situe à l'extérieur de ladite limite admissible
de valeur.
16. Procédé selon la revendication 1, dans lequel ladite étape de réglage de la valeur
limite consiste à substituer une composante caractéristique limite à ladite pluralité
de composantes caractéristiques, lorsque lesdites composantes caractéristiques sont
inférieures à la composante caractéristique limite.
17. Procédé selon la revendication 1, dans lequel ladite étape de réglage de la valeur
limite consiste à substituer une composante caractéristique limite devant être utilisée
en commun pour n'importe laquelle de ladite pluralité de composantes caractéristiques
lorsque ladite composante quelconque parmi la pluralité de composantes caractéristiques
devient inférieure à ladite composante caractéristique limite.
18. Procédé selon la revendication 1, selon lequel une évaluation est faite pour savoir
si l'ensemble de ladite pluralité de composantes caractéristiques est inférieur ou
non à une composante caractéristique limite respective, qui lui correspond ; et
on substitue ladite composante caractéristique limite respective, qui lui correspond,
à l'une quelconque d'une pluralité de composantes caractéristiques, qui est évaluée
comme étant inférieure à ladite composante caractéristique limite respective.
19. Procédé de commande de séquence d'allumage comprenant les étapes consistant à:
extraire une pluralité de composantes caractéristiques (fi) indicatives d'un cognement
à partir de quantités physiques (BGLi) concernant un état de cognement dans un moteur
à combustion interne, et une pluralité de composantes de comparaison devant être définies
conformément auxdites quantités physiques ;
extraire une information concernant l'apparition de cognement (107, 108, 112) à partir
d'une relation (106) entre ladite pluralité de composantes caractéristiques et ladite
pluralité de composantes de comparaison ;
régler une séquence d'allumage de manière à réduire un cognement lors de l'apparition
de ce dernier ;
mettre à jour l'une desdites composantes de comparaison lorsqu'il n'existe aucun cognement,
avec une nouvelle valeur qui est située dans une gamme de valeurs que l'on peut prendre
pour ladite composante de comparai-son ; et
mettre à jour l'une desdites composantes de comparaison avec une composante de comparaison
limite prédéterminée lorsque ladite composante de comparaison est située de dehors
de la gamme admissible de valeurs de cette composante.
20. Commande de séquence d'allumage selon la revendication 19, dans laquelle lesdites
étapes de réglage et de mise à jour comprennent les étapes consistant à
évaluer l'apparition d'un cognement lorsqu'une somme d'au moins deux desdits rapports
de composantes devient supérieure à une valeur prédéterminée ;
régler la séquence d'allumage de manière à réduire le cognement lors de son apparition
lorsqu'il est évalué comme étant apparu comme indiqué précédemment ;
déterminer, lors de l'évaluation du fait qu'il n'existe aucun cognement lorsque la
somme desdits au moins deux rapports de composantes est inférieure à ladite valeur
prédéterminée, si lesdites composantes de comparaison respectives sont supérieures
à des composantes de comparaison limites respectives qui leur correspondent, ou sinon
à une composante de comparaison limite sélectionnée de façon particulière ;
mettre à jour lesdites composantes de comparaison respectives avec de nouvelles valeurs
respectives lorsqu'elles sont supérieures à la composante de comparaison limite ;
et
mettre à jour lesdites composantes de comparaison respectives avec lesdites composantes
respectives de comparaison limites qui leur correspondent ou avec ladite composante
de comparaison limite lorsque lesdites composantes de comparaison respectives sont
inférieures à la ou aux composantes de comparaison limites.